© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 5846
    Prev
    Next
    INDEX
    Explanations

    possible loss / most efficient / potential sentiment

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires most strongly on very common, high‐frequency English tokens—especially short function words (e.g. “to,” “no,” “always,” “multiple”) and other frequent words—i.e. it flags high‐frequency vocabulary.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     отмечает
    0.84
    早速
    0.82
    Estamos
    0.79
     চার
    0.78
    Những
    0.77
     Estamos
    0.75
     शुक्रवार
    0.75
    近年来
    0.74
    ➢
    0.74
    UPDATE
    0.73
    POSITIVE LOGITS
     photons
    0.79
     molecules
    0.71
     electrons
    0.69
     melakukannya
    0.68
     other
    0.66
     neg
    0.66
     bytes
    0.64
     pronouns
    0.62
     akzept
    0.61
     drugih
    0.60
    Activations Density 0.957%

    No Known Activations